Table 3. Multivariate models.

Predictor		η2	F	Effect df	Error df	p
Model 1: All growth and anthropometric variables; sucrose and salt preference as joint outcome measures
Sex		0.07	2.52	2	68	0.088
Height (m)		0.14	5.57	2	68	0.006
NTx/creatinine		0.12	4.74	2	68	0.012
Percent body fat		0.12	4.42	2	68	0.016
Age (years)		0.05	1.73	2	68	0.186
Model 2: Same as Model 1 but with no NTx outliers
Sex		0.09	3.40	2	64	0.040
Height (m)		0.15	5.74	2	64	0.005
NTx/creatinine		0.01	0.19	2	64	0.826
Percent body fat		0.11	3.91	2	64	0.024
Age (years)		0.06	2.03	2	64	0.140
Model 3: Same as Model 2 but with sucrose preference as only outcome measure
Sex		0.07	5.33	1	65	0.024
Height (m)		0.15	11.43	1	65	0.001
NTx/creatinine		0.00	0.29	1	65	0.588
Percent body fat		0.02	1.29	1	65	0.260
Age (years)		0.05	3.45	1	65	0.068
Model 4: Same as Model 2 but with salt preference as only outcome measure
Sex		0.03	2.30	1	65	0.13
Height (m)		0.01	0.78	1	65	0.38
NTx/creatinine		0.00	0.14	1	65	0.71
Percent body fat		0.08	5.89	1	65	0.02
Age (years)		0.02	1.08	1	65	0.30

η², proportion of variance accounted for; F, Fisher value; df, degrees of freedom. Only children with complete data for these variables were included in the analysis (e.g., n = 78 for Model 1). Variables shown in boldface made a significant contribution to the model.